NVIDIA GeForce GTX 670M vs NVIDIA GeForce GTX 675M Comparison
NVIDIA GeForce GTX 670M
GeForce GTX 675M
PERFORMANCE BENCHMARKS
Analysis: NVIDIA GeForce GTX 670M vs NVIDIA GeForce GTX 675M
The NVIDIA GeForce GTX 675M and NVIDIA GeForce GTX 670M are both Fermi 2.0-era mobile graphics processors from the GeForce 600M generation, built on TSMC's 40 nm process with 1,950 million transistors on a 332 mm² die. The benchmark data places the GTX 675M ahead, with a Geekbench OpenCL score of 6946 versus 6513 for the GTX 670M, a 6.6% advantage. Both cards are end-of-life products released on the same date, but the GTX 675M consistently outperforms the GTX 670M across the single available benchmark, making it the stronger choice for compute-heavy workloads.
FAQ
Q: What is the performance difference between the GTX 675M and GTX 670M in the available benchmark?
A: In the Geekbench OpenCL test, the GTX 675M scores 6946, while the GTX 670M scores 6513. This gives the GTX 675M a 6.6% higher score, making it the clear winner in this head-to-head comparison.
Q: How does each card rank against all other GPUs in the database?
A: The GTX 675M sits in the 39th percentile of all GPUs, while the GTX 670M sits in the 38th percentile. This indicates both are mid-range performers, but the GTX 675M holds a slight edge in overall standing.
Q: What are the closest rivals for the GTX 675M?
A: The nearest rivals to the GTX 675M are the AMD Radeon R5 M240 (avg score 6975, 0.4% lower), the NVIDIA GeForce GTX 680M (avg score 7023, 1.1% lower), the NVIDIA T600 (avg score 7035, 1.3% lower), and the AMD FirePro M5100 (avg score 6830, 1.7% higher).
Q: What are the closest rivals for the GTX 670M?
A: The nearest rivals to the GTX 670M are the NVIDIA GeForce GT 555M (avg score 6493, 0.3% higher), the NVIDIA Quadro M5000M (avg score 6481, 0.5% higher), the AMD Radeon Vega 10 Mobile (avg score 6476, 0.6% higher), and the Intel UHD Graphics P750 (avg score 6554, 0.6% lower).
Q: Do the two cards share the same memory clock speed?
A: Yes, both the GTX 675M and GTX 670M have a memory clock of 750 MHz, which translates to 3 Gbps effective. However, the memory configuration differs, with the GTX 675M using a 256-bit bus and the GTX 670M using a 192-bit bus.
Q: Are both cards from the same architecture and process node?
A: Yes, both are built on the Fermi 2.0 architecture using TSMC's 40 nm process, with the same GF114 chip. They share identical transistor counts (1,950 million) and die size (332 mm²), as well as the same transistor density of 5.9M / mm².
Where Each One Wins
The GTX 675M wins in every available benchmark category, with a 6.6% lead in Geekbench OpenCL. This makes it the superior choice for general-purpose compute tasks, OpenCL acceleration, and any workload that leverages the GPU's shading units. The GTX 675M's advantage stems from its higher core configuration: it packs 384 shading units, 64 texture mapping units, and 32 raster operations pipelines, compared to the GTX 670M's 336 shading units, 56 TMUs, and 24 ROPs.
The GTX 670M, while behind in raw compute, does have one notable advantage: power consumption. Its TDP is 75 W, significantly lower than the GTX 675M's 100 W. This makes the GTX 670M a more power-efficient option for thin-and-light laptops where thermal headroom and battery life are priorities. The GTX 670M also has a smaller memory footprint at 1536 MB, which could be sufficient for less demanding applications while reducing cost in OEM configurations.
In terms of percentile rank, the GTX 675M's 39th percentile versus the GTX 670M's 38th percentile is a narrow margin, but the actual benchmark delta is more pronounced. The GTX 675M is the pick for users who need maximum compute throughput per frame or per task, while the GTX 670M is the pick for those who prioritize lower power draw without a drastic performance drop.
Architecture Differences
Both GPUs share the same fundamental architecture: Fermi 2.0 on TSMC's 40 nm process, with the GF114 chip. The transistor count is identical at 1,950 million, and the die size is the same at 332 mm², yielding a transistor density of 5.9M / mm². The memory type is also the same (GDDR5), and the memory clock is identical at 750 MHz (3 Gbps effective). Both cards expose the same API support: DirectX 12 (11_0), OpenGL 4.6, and no Vulkan support. They also share the same bus interface (MXM-B 3.0), slot width (MXM Module), and display outputs (Portable Device Dependent).
The core architectural difference lies in the execution resources. The GTX 675M has 384 shading units, 64 TMUs, and 32 ROPs, while the GTX 670M has 336 shading units, 56 TMUs, and 24 ROPs. This translates to higher pixel and texture fill rates for the GTX 675M: 9.920 GPixel/s and 39.68 GTexel/s, respectively, versus 8.372 GPixel/s and 33.49 GTexel/s for the GTX 670M. The FP32 performance also scales accordingly: 952.3 GFLOPS for the GTX 675M versus 803.7 GFLOPS for the GTX 670M.
The memory subsystem is where the second major difference appears. The GTX 675M uses a 256-bit memory bus with 2 GB of GDDR5, yielding a bandwidth of 96.00 GB/s. The GTX 670M uses a narrower 192-bit bus with 1536 MB of GDDR5, capping bandwidth at 72.00 GB/s. This 33% bandwidth advantage for the GTX 675M is critical for memory-intensive workloads like high-resolution textures or compute kernels that stream data.
Specification Differences
The GTX 675M and GTX 670M differ in several key specifications. The GTX 675M has a higher shading unit count (384 vs. 336), more TMUs (64 vs. 56), and more ROPs (32 vs. 24). Its pixel rate is 9.920 GPixel/s versus 8.372 GPixel/s, and its texture rate is 39.68 GTexel/s versus 33.49 GTexel/s. FP32 performance is 952.3 GFLOPS for the GTX 675M, compared to 803.7 GFLOPS for the GTX 670M.
Memory configuration also diverges: the GTX 675M has 2 GB of GDDR5 on a 256-bit bus, delivering 96.00 GB/s of bandwidth, while the GTX 670M has 1536 MB on a 192-bit bus, delivering 72.00 GB/s. The TDP is another differentiator, with the GTX 675M rated at 100 W and the GTX 670M at 75 W. Both cards share the same process node, chip, architecture, transistor count, die size, memory clock, API support, bus interface, slot width, power connectors (None), display outputs, and release date.
Head-to-Head Benchmarks
The only head-to-head benchmark available is Geekbench OpenCL, and the GTX 675M wins decisively. It scores 6946, while the GTX 670M scores 6513, resulting in a delta of 6.6% in favor of the GTX 675M. This is a substantial margin for two cards from the same generation and chip family.
Breaking down what drives this delta, the GTX 675M's extra 48 shading units (384 vs. 336) directly boost FP32 throughput, which is the primary workload in OpenCL compute tests. The 25% increase in TMUs (64 vs. 56) and 33% increase in ROPs (32 vs. 24) also contribute to higher fill rates, which can matter in certain compute workloads that involve texture sampling or rasterization. The memory bandwidth advantage is arguably the most impactful: 96.00 GB/s versus 72.00 GB/s is a 33% gap, which can be the bottleneck for memory-bound kernels.
In the context of nearest rivals, the GTX 675M's score of 6946 places it just 0.4% behind the AMD Radeon R5 M240 (6975) and 1.1% behind the NVIDIA GeForce GTX 680M (7023). It is 1.3% behind the NVIDIA T600 (7035) but 1.7% ahead of the AMD FirePro M5100 (6830). The GTX 670M's score of 6513 puts it 0.3% ahead of the NVIDIA GeForce GT 555M (6493), 0.5% ahead of the NVIDIA Quadro M5000M (6481), 0.6% ahead of the AMD Radeon Vega 10 Mobile (6476), and 0.6% behind the Intel UHD Graphics P750 (6554). These comparisons show that both cards are tightly clustered with their peers, but the GTX 675M sits in a higher performance band.
The Verdict
The data is unambiguous: the NVIDIA GeForce GTX 675M is the faster GPU. It wins the only available head-to-head benchmark with a 6.6% margin, and its percentile rank (39th vs. 38th) reflects a slightly higher standing in the overall GPU hierarchy. For any user prioritizing raw compute performance, OpenCL workloads, or gaming at higher resolutions and detail settings, the GTX 675M is the correct choice.
The GTX 670M is not without merit. Its 75 W TDP versus the GTX 675M's 100 W makes it a better fit for systems where power consumption and thermal output are critical constraints. For a laptop chassis with limited cooling, the GTX 670M could sustain longer boost behavior or run quieter. Its 1536 MB frame buffer is smaller, but for less demanding titles or compute tasks, this may not be a limiting factor.
However, the performance gap is not trivial. A 6.6% delta in OpenCL translates to visible differences in compute tasks, and the GTX 675M's 33% higher memory bandwidth (96.00 GB/s vs. 72.00 GB/s) means the gap could widen further in memory-bound scenarios. If the choice is between these two cards and power draw is not an overriding concern, the GTX 675M is the definitive pick. If the system demands a lower TDP, the GTX 670M offers a respectable fallback, but it concedes the performance crown in every measurable way.